We are testing the failover behaviour of jboss esb. We are using jboss 5.1 GA, jboss ESB 4.6, a clustered post office, clustered queues, with oracle as a db.
We have noticed the following behaviour. When we fail on purpose one node of our ESB (we kill the java process using task manager in windows) the messaging on the alive node starts acting up for 2 minutes before the system goes back to normal. After the 2 minutes the system works like nothing has happened.
We were wondering what all the errors we see in our logs are (I attach a file) and whether there is a possibility through configuration to get this time to something smaller than 10 seconds.